1.0.2 - Rocket ascent profile and orbit delta-V

[Temeter]

I'd say 3400m/s dV looks to be around the minimum - but I'd love to be proved wrong

Recently did an ascent with 3290 dv, and that was purely by accident. One thing to note was the low trajectory and a relatively strong 2nd stage (rocket was 2 stage + booster) with 1.55 t/w.

[Slugy]

Recently did an ascent with 3290 dv, and that was purely by accident. One thing to note was the low trajectory and a relatively strong 2nd stage (rocket was 2 stage + booster) with 1.55 t/w.

You beat me too it: I was just going to edit my post to say that was for a high payload launch (~22%), and that 3300/s looked likely for lower ones.

Looks like better is possible, if you got under that without making a special effort!

[soulsource]

As fascinating as all this is (and it's been helpful for sure), I am left with questions about optimal speed during ascent to avoid creating excessive drag forces. Previously, I would follow the terminal velocity values, and I've continued doing that into 1.0, but this often results in my going hypersonic (flames) on the way up. Am I going way too fast? Or perhaps it's not such a huge deal anymore and I should just put the hammer down and get to orbit as fast as I can without making anything explode? The rocket I've been doing most of the playing with since the patch can pull 3+ g's the entire way to establishing apoapsis.

I also have to say the more pronounced gravity turn is a bit freaky. It really feels like I'm about to nose dive back into the planet. It's nice to be so close to a circular orbit though from the initial burn. Way less work to do in space, and the short burn creates a more even orbit.

The math showing that for a vertical ascent going at terminal velocity is optimal at any given point is sound and still applies. In my opinion there is nothing wrong with a hypersonic ascent. Nevertheless, I've got the impression that one often reaches limits in stability and controllability before terminal velocity, so I've basically given up chasing for it, and just set my initial TWR of the first stage to about 1.5...

[ColKlonk]

I'm with VJeb on the ascent profile.. Although profiles vary according to rocket mass - heavier = more vertical before the turn.

I made an Exel drag profile (needs checking) sheet using NASA's atmospheric density formulae (http://exploration.grc.nasa.gov/education/rocket/atmos.html), and found that going high first(approx 10% of atmosphere height), before the turn results equal or less (mostly) drag, than turning early.

My logic here is that you have to get to a certain orbital height and have to overcome this G-Drag.. and this is constant no matter what profile you use.

What matters is how you get there, and the biggest limiting factor is Atmospheric Drag.

By turning early you spend more time in the atmosphere = more drag. Not only this.. you have to go faster in order to keep your nose up = A lot lot more drag.

By turning late, you are for the most part 'on prograde' =less drag + more control, and can travel slower (at approx 1.25Gs) = less drag. This also gives you the luxury of a wider error window, which is easier to correct.

Once higher up.. you can start with the 'after-burners', but if you get it right (I still need practise )this might not be necessary.

Here's a 4x speed vid on a RSS ascent. 2,000 tons including a 70-80 ton payload. The same method would apply to the Stock system

https://www.dropbox.com/s/duyo5plqvbkwunt/Rock2000_1.mp4?dl=0

Edited May 5, 2015 by ColKlonk

[Temeter]

The math showing that for a vertical ascent going at terminal velocity is optimal at any given point is sound and still applies. In my opinion there is nothing wrong with a hypersonic ascent. Nevertheless, I've got the impression that one often reaches limits in stability and controllability before terminal velocity, so I've basically given up chasing for it, and just set my initial TWR of the first stage to about 1.5...

The vertical ascent is very inefficient in terms your of flight path, tho. Transforming the ascent into the shallower trajectory is hard at high speeds and can cause a lot of drag losses. Low speeds on the other hand cause less ISP and more gravity losses.

Edited May 5, 2015 by Temeter

[eviator]

Well I tried Violent Jeb's profile and came across some issues.

#1 going from near vertical at 7km to 30 degrees from horizontal at around 20km requires me to burn well off prograde, which increases drag. Should I maintain reduced thrust to maintain this?

#2 I cut my engines around 50k as maintaining 1 minute to Ap puts me into the red on the navball. My Ap is still inside the atmosphere. So buring horizontally within 10 seconds to Ap causes me to flatten my trajectory while still inside the atmosphere. Eventually I'm going fast enough to flip my Ap to the other side of the planet. But since I'm still in the atmosphere with a flat trajectory, the small amount of drag is affective for several minutes. My Ap eventually drops back into the atmosphere, and the process continues. So with your suggested profile, I'm spending most of the time trying to maintain orbit inside the atmosphere.

To avoid #2, I'd have to make sure my Ap is a margin outside the atmosphere before cutting engines. This means either the time to Ap has to go beyond 1 minute, or I need a more vertical profile. Either way, it does not seem like the profile you described works all that well.

[Randox]

The math showing that for a vertical ascent going at terminal velocity is optimal at any given point is sound and still applies. In my opinion there is nothing wrong with a hypersonic ascent. Nevertheless, I've got the impression that one often reaches limits in stability and controllability before terminal velocity, so I've basically given up chasing for it, and just set my initial TWR of the first stage to about 1.5...

Good to know the math still works out. I am, of course, not ascending vertically. At some point I'll get around to actually tracking my dV expenditure and possibly fine tuning it. Based on stuff I've been reading, and the limits on how fast I can roll the rocket into a turn, I've been trying to restrain my speed (likely not enough) in the lower atmosphere, and wait until I am in thinner air and closer to level to really gun the engines. It's working quite well in so far as my circularization burn at apoapsis is usually 200-400 m/s, instead of my customary 1100-1300.

Can't say I've had any stability problems related to aerodynamics with the rockets. Both of the ones I've been making use of since 1.0 are quite happy to go transonic. If anything, the heavy lifter in particular enjoys the extra control authority granted to the fins.

I also feel like my TWR on...everything, might be excessive. However, since I have a rocket that can blow past terminal velocity like it's being chased by angry ghosts, I might do some test launches and see if speed ever causes control issues.

Edited May 5, 2015 by Randox

[ColKlonk]

Well I tried Violent Jeb's profile and came across some issues.

#1 going from near vertical at 7km to 30 degrees from horizontal at around 20km requires me to burn well off prograde, which increases drag. Should I maintain reduced thrust to maintain this?

#2 I cut my engines around 50k as maintaining 1 minute to Ap puts me into the red on the navball. My Ap is still inside the atmosphere. So buring horizontally within 10 seconds to Ap causes me to flatten my trajectory while still inside the atmosphere. Eventually I'm going fast enough to flip my Ap to the other side of the planet. But since I'm still in the atmosphere with a flat trajectory, the small amount of drag is affective for several minutes. My Ap eventually drops back into the atmosphere, and the process continues. So with your suggested profile, I'm spending most of the time trying to maintain orbit inside the atmosphere.

To avoid #2, I'd have to make sure my Ap is a margin outside the atmosphere before cutting engines. This means either the time to Ap has to go beyond 1 minute, or I need a more vertical profile. Either way, it does not seem like the profile you described works all that well.

You're probably turning to fast... You should come out the atmosphere at about 20-30 degrees with your AP anywhere between 60-90 seconds ahead. If you lock onto the Orbit Prograde marker on the way up (at approx half atmosphere height), you can control your AP position via throttle control. I haven't played stock Kerbin for a while, but I think the atmosphere limit is around 55km.. you should aim your AP at around 80-100km.. and thus your orbit.

Like gravity, and other turns, it takes a bit of practice to perfect... and it changes with each different rocket.

Also there is no rush to get to AP. Find out what the required orbital velocity is and you can gently cruise to this, by keeping AP a few seconds ahead, letting AP approach you as you near orbital velocity.

Test a lot of different rocket designs, and eventually you'll settle on a few reliable variants. You get to know each rocket stage parameter (eg: Height, angle and velocity at stage changes) that is required for successful orbit

During testing of each stage of my rocket designs, I just launched the stage vertically and it had to get out of the atmosphere straight up. Plonked the stages together with a payload and 'fine tuned' the final design.

After going through this lot, Stage-1 had to get you to half atmosphere height, Stage-2 into pre-orbit (or orbit), and Stage-3 into orbit and beyond.

With #2, a solution would be to turn the rocket near vertical (pointing straight up) and give the engines a good boost... this should help you out of the atmosphere.

Edited May 6, 2015 by ColKlonk

[eviator]

You're probably turning to fast...

Well I appreciate the long response, but it's not quite what I'm looking for. I can get into orbit just fine, I'm trying to find some good rules for optimal (minimum dV) ascent profiles, thus why I tried Violent Jeb's suggestion.

My average good orbit is about 3550 vacuum dV. I've tweaked my stage TWRs, I've done the SAS on prograde gravity turn, I've done the no SAS gravity turn. I've minimized drag with fairings, I've done shallow ascents (horizontal by 20km), I've done steep ascents (no gravity turn until 7k), and everything in between. Because folks have been able to get orbits with 250-300m/s less than my best, there is some key that I'm missing.

[Gaugeforever]

Well I appreciate the long response, but it's not quite what I'm looking for. I can get into orbit just fine, I'm trying to find some good rules for optimal (minimum dV) ascent profiles, thus why I tried Violent Jeb's suggestion.

My average good orbit is about 3550 vacuum dV. I've tweaked my stage TWRs, I've done the SAS on prograde gravity turn, I've done the no SAS gravity turn. I've minimized drag with fairings, I've done shallow ascents (horizontal by 20km), I've done steep ascents (no gravity turn until 7k), and everything in between. Because folks have been able to get orbits with 250-300m/s less than my best, there is some key that I'm missing.

If I had to guess I would say it's drag that you are missing. You probably have some items on your craft that have drag values that are slowing down your accent in atmo. You need that sleek aero rocket to shave some DV.

[eviator]

If I had to guess I would say it's drag that you are missing. You probably have some items on your craft that have drag values that are slowing down your accent in atmo. You need that sleek aero rocket to shave some DV.

That's a good guess. What part of this rocket is creating extra drag?

[Foxster]

You can turn on the numbers for drag on each part in the debug menu under Physics. I'd say though the biggest drag will be from the shell and the adapter.

[Deutherius]

I'm pretty sure that burning at AP/PE is only universally efficient when you're already in orbit and have no other considerations. If that was true in all circumstances, we'd be pointing straight up and circularizing at the top of a vertical trajectory.

You have picked my comment out of context - I was reacting to severedsolo's comment, in which he described a pretty normal gravity turn and ascent to orbit. In that case, it will be most optimal to circularize at Ap. If he wants a lower orbit then the 100 km he described, it's going to be better to raise his Ap only to that altitude (instead of pushing Ap too far and then burning radial-in to compensate).

[Violent Jeb]

That's a good guess. What part of this rocket is creating extra drag?

http://i1295.photobucket.com/albums/b621/gotyoke/KSP/2015-05-06_00002_zps3aofnfzp.jpg

I think i see why you are unable to match my profile.

Your craft is massively powerful. The thrust limiter in the VAB is essentially the same as the reducing the throttle, the engine still has the same efficiency, so when you are lowering your max thrust, you are carrying the extra weight of your engine up with you. That is where the 300-500 difference is.

The total time I am throttled down is less than 30s, only during the thickest part of the atmosphere where i'm losing too much to the wind (your design is throttled 60% the entire time). Drag is seriously reduced approaching 20km (inside the second layer of atmo), so I throttle it back up to 100%

As it is, your craft is able to blast way beyond the Ap by the time you're ready to start the turn, which means you need even less throttle from the engine.

I would try either an additional fuel can on your initial stage, or a smaller engine, so that you don't rocket past (a good pun) that 1minute Ap. The additional fuel can will add to your dV while making your launch more efficient - so i'd go that way. Double the bottom stage fuel throttle 100% and you should see those savings

Edited May 6, 2015 by Violent Jeb

[NikkyD]

Ok, im calling BS on the 3.000 dV to LKO. I guess some control surfaces generate lift because he is turning so extreme and has more of a 90% rocket 10% plane.

I just did a lot of profiles and roughly 3500 dV is the figure.

I use a really simple design, MK16 parachute on top of a MK1 cockpit without monoprop, service bay to store mechjeb in (just in case, avoid drag), followed by 2x4.5t tanks and one lv30 engine.

The engine has 280-300 ISP so my dV estimate atm/vac is roughly the same.

This setup gives me 3414/3658 dV and more than 100 dV(vac) left i couldnt pull off. Maybe 100 dV less, maybe 150 but no way 500+.

I usually play with the "all realism" mods because stock is very childish but now with fairing and realistic drag i gave it a shot.

Terminal Velocity should still be valid, so i don't know why you all are going so slow with 1,5 TWR and doing kamikaze turns.

If you do that, you need to find the sweetspot between drag and gravity flipping your rocket over but you will still be within the atmosphere way too long if you finish the turns at 20km.

Using winglets to be able to actually fly your maneuvers already shows that something is wrong!

[eviator]

Ok, im calling BS on the 3.000 dV to LKO. I guess some control surfaces generate lift because he is turning so extreme and has more of a 90% rocket 10% plane.

I just did a lot of profiles and roughly 3500 dV is the figure.

Yeah, I'm to this point as well. Even 3,300 seems unlikely to me and unfortunately nobody is willing to share their design with detailed profile. My average best is in the 3,500 range with a streamlined rocket and essentially no payload. When trying to get stuff into orbit, I have suboptimal drag and staging, so I'm depending upon 3,800 dV to get to LKO.

[herbal space program]

I haven't made any actual deltaV calculations, but based on doing it again and again and again I'd say my most efficient orbits in the current system have involved accelerating to around 150 m/s as quickly and vertically as I possibly can, then throttling back significantly and starting to go off vertical a tiny bit at around 3,000m, keeping my heading inside the prograde marker but not quite in the middle of it, so that I hit 45 degrees at around 10km and 22.5 degrees around 20 km. I then boost completely prograde until maybe 35 km and then drop my heading to the horizon, cutting thrust at around 45 km. Usually I only need about another 300 or so m/s to circularize at a 70km apoapsis. Speedwise, as I said I go straight to around 150 m/s, then slow my acceleration so I'm doing maybe 300 m/s at 10km, 600 at 20km, and 1200 at 30 km.

Edited May 9, 2015 by herbal space program

[Temeter]

Ok, im calling BS on the 3.000 dV to LKO. I guess some control surfaces generate lift because he is turning so extreme and has more of a 90% rocket 10% plane.

I just did a lot of profiles and roughly 3500 dV is the figure.

I don't think you can judge that when your best ascent is is above 3400 dV. People managed 2800 dV in 1.0, so 3k in 102 shouldn't come as a surprise.

[NikkyD]

I don't think you can judge that when your best ascent is is above 3400 dV. People managed 2800 dV in 1.0, so 3k in 102 shouldn't come as a surprise.

I skipped 1.0 + 1.0.1 but had every one before that and i don't see anywhere under 3.4 without exploits or semi planes and in every version i was trying to find the most efficient ascent, so please...

further 1.0.2 tests:

- going to 30 and speeding with 0° to reach AP of 80 isnt working well. Still too low, i see hot-atmo effect before 2km/s, fuel runs out

- works at 35 but not very efficient

- at 35 aiming 10° works better, requires circ.burn of about 400 dV. Reaching AP of 80 at about 40km alt with 2000 km/s, losing up to 100 m/s while coasting

- old style, 5/70 @100% (mechjeb) i reach AP too quickly with only 600 m/s orbital, ends in about 3.9k dV

best attempt is a 1/35@100% to 5° profile with an accel-limit of about 20. Can speed up more with gimbal engines, might not make it without gimbal depending on design.

winglets cause more drag than they are worth.

parachute or aero nosecone doesnt make any difference.

rocket may flip backwards if very long and fuel empties out from top, so i had to make fuel go back up to keep the nose heavy.

Main problem is stability vs speed.

- If you go too slow you spend too much time in atmo

- If you are REALLY slow the nose might fall to the ground

- too fast means course change happens too quickly and drag is pushing the nose down -> spinning

I dont see where 500 of 3500 might be saved

[Temeter]

I skipped 1.0 + 1.0.1 but had every one before that and i don't see anywhere under 3.4 without exploits or semi planes and in every version i was trying to find the most efficient ascent, so please...

It seems like you weren't very succesfull.

further 1.0.2 tests:

- going to 30 and speeding with 0° to reach AP of 80 isnt working well. Still too low, i see hot-atmo effect before 2km/s, fuel runs out

- works at 35 but not very efficient

- at 35 aiming 10° works better, requires circ.burn of about 400 dV. Reaching AP of 80 at about 40km alt with 2000 km/s, losing up to 100 m/s while coasting

- old style, 5/70 @100% (mechjeb) i reach AP too quickly with only 600 m/s orbital, ends in about 3.9k dV

best attempt is a 1/35@100% to 5° profile with an accel-limit of about 20. Can speed up more with gimbal engines, might not make it without gimbal depending on design.

winglets cause more drag than they are worth.

parachute or aero nosecone doesnt make any difference.

rocket may flip backwards if very long and fuel empties out from top, so i had to make fuel go back up to keep the nose heavy.

Main problem is stability vs speed.

- If you go too slow you spend too much time in atmo

- If you are REALLY slow the nose might fall to the ground

- too fast means course change happens too quickly and drag is pushing the nose down -> spinning

I dont see where 500 of 3500 might be saved

Many good trajectories need very little course corrections, gravity and aerodynamics will efficiently move the apoapsis, which allows the rocket to always stay directly in the air stream and minimizing drag. The really efficient launchs seem to go for very low trajectories, and rely on high speed to push a relatively distant apoapsis up.

Btw, I managed to get 3290 by instantly turning the a rocket between 5 and 10°, using stability assist until ~1km, and locked to prograde. Then just burned until the apoapsis was at 80km without any input and circularized later. It can be this simple. Craft was a bog standard 2.5m rocket, afair with some dummy cargo.

Edited May 9, 2015 by Temeter

[cybersol]

I haven't made any actual deltaV calculations, but based on doing it again and again and again I'd say my most efficient orbits in the current system have involved accelerating to around 150 m/s as quickly and vertically as I possibly can, then throttling back significantly and starting to go off vertical a tiny bit at around 3,000m, keeping my heading inside the prograde marker but not quite in the middle of it, so that I hit 45 degrees at around 10km and 22.5 degrees around 20 km. I then boost completely prograde until maybe 35 km and then drop my heading to the horizon, cutting thrust at around 45 km. Usually I only need about another 300 or so m/s to circularize at a 70km apoapsis. Speedwise, as I said I go straight to around 150 m/s, then slow my acceleration so I'm doing maybe 300 m/s at 10km, 600 at 20km, and 1200 at 30 km.

I fly manually and find my optimal ascent profile is very similar to this. My best profiles are slightly higher, hitting 45* at around 12 km. Also, I don't generally find controlling speed to really be needed. With 1.8+ start TWR this profile can reliably achieve 3300-3350 dV for me.

[NikkyD]

Are you (all) talking atmo or vac dV, cause i only look at vac

[cybersol]

Are you (all) talking atmo or vac dV, cause i only look at vac

I was reporting vac dV expended. Also, I tried your simple rocket design you posted earlier. Its obvious you were not flying that manually with a keyboard -- it cannot get even halfway to the edge of the prograde marker without spinning. To fly it manually, I needed to swap the service bay for a 1.25m advanced inline stabilizer and add 3 av-t1 winglets. With those changes I used 3370 vac dV on my first attempt to LKO.

[NikkyD]

Its obvious you were not flying that manually with a keyboard

Ofc i let MJ do the ascent, the only way its accurate and reproduceable.

Will try your profile tomorrow, getting kinda late for me

but i have to say if you have to use winglets afaik something is wrong because a rocket shouldnt have winglets

[LostOblivion]

Especially now that thrust also varies with atm. pressure, and mostly because of the rather large difference in engine specific impulse now, you really can't just measure delta-v accurately by looking at the vac. delta-v before and after, but you have to actually measure the aggregate expended delta-v throughout the launch. There are mods for this.

This also goes for TWR, now more than ever before. I wish KER made the switch for atm. sea level quickly available in the compact view now that it's so important.